Construct Validity of Animal-Assisted Therapy and Activities: How Important Is the Animal in AAT?

Animal-assisted therapy and animal-assisted activities involve a nonhuman animal as a key therapeutic agent in some kind of intervention that may range from highly specified, as in AAT, to more casual, as in AAA. In this review I address the question: How important is the animal in animal therapy? In other words, does the recent literature strongly support the notion that a live animal, as opposed to another novel stimulating component, is specifically necessary for therapeutic success. Two meta-analyses and 28 single empirical studies were reviewed in order to address this issue. I conclude that the effects of AAT and AAA are likely to be moderate and broad at best and that, although improving, the literature has not yet reached an experimentally rigorous enough level to provide a definitive robust conclusion about the effectiveness of these approaches, particularly with regard to the question of whether a live animal is necessary for a therapeutic effect. Animal-assisted activities (AAA) and animal-assisted therapy (AAT) are becoming increasingly popular adjuncts for more traditional forms of therapy in a variety of settings, such as schools, hospitals, and psychiatric centers. The range of nonhuman animals (hereafter referred to as “animals”) used and the types of maladies targeted are expansive. Whereas domesticated animals, particularly dogs and horses, are the most popular choice (Fine 2010), there are many facilities and programs that offer “therapy” with less commonly used animals such as fish (Cole and Galinski 1995, 2000) and birds (Davis 1988), as well as exotic animals such as dolphins (Marino and Lilienfeld 1998, 2007) and even elephants (Swanepoel and Odendaal 2005). Disorders targeted range from generalized psychological problems to depression and anxiety to developmental disabilities and infectious disease (Fine 2010). Various terms have been used to describe AAA and AAT such as pet therapy, petfacilitated therapy, pet-assisted therapy, and animal-facilitated therapy (Connor and Miller 2000). Pet Partners (formerly Delta Society), a leading international, non-profit organization that provides training for AAA and AAT practice, defines AAA and AAT in the following manner: AAA provide opportunities for motivational, educational, recreational, and/or therapeutic benefits to enhance quality of life. AAA is delivered in a variety of environments by specially trained professionals, paraprofessionals, and/or volunteers, in association with animals that meet specific criteria. (“Standards of Practice” 1996) AAT is a goal-directed intervention in which an animal that meets specific criteria is an integral part of the treatment process. AAT is directed and/or delivered by a health/human service provider working within the scope of practice of his/her profession. (“Standards of Practice” 1996) AAT is distinguishable from AAA in that it has a specific therapeutic objective and is delivered as part of a highly specific treatment plan. AAA, on the other hand, involves a more casual interaction between animals and humans and can be provided by a broader range of individuals than AAT. Although Pet Partners distinguishes AAA and AAT, in actual practice the two are often conflated and may overlap considerably. In AAT and AAA the animal serves a variety of roles from companionship to caregiver to social facilitator and reinforcement; studies often do not provide enough information about the way the animal was used to differentiate between AAT and AAA. Despite this ambiguity, Pet Partners is clear about the central point that although often used as an adjunct to other forms of therapy that include the presence of a human, AAT and AAA are based on the premise that the animal makes a therapeutic difference to the subject. The purported mechanisms of action include mental stimulation, physical stimulation, physiological effects, and general motivating effects. Therefore, according to Pet Partner’s claims, the effects of an animal in the presence of a human on the subject should be differentiable from the effects of just a human on the subject and differentiable from the effects of another novel stimulus and a human. The implication of animal-assisted interventions of all kinds is that the live animal is a highly specific component of the therapy. That is, without the animal there is no “animal-assisted” activity of any sort. The entire industry is founded on the premise that involvement of a live animal in any kind of activity or therapeutic intervention (no matter how complex or multidimensional) is critical to its identity. Therefore, it is important to determine efficacy and specificity, that is, whether AAT and or AAA result in real therapeutic improvement and, if so, whether the improvement is due to the presence of the animal specifically. Hence, the focus of this review and evaluation is to address the question: How important is the animal in animal-assisted therapy? There is a longstanding general sense among practitioners and researchers that activities and interventions involving animals may have nonspecific positive effects on both patients and normative populations of subjects. Yet, there have also been lingering doubts and controversies about the robustness and specificity of the effect of animals on human health and welfare. In other words, there remain questions about the validity of AAT and AAA. In a 2002 review of the field, Johnson, Odendaal and Meadows recognized the importance of obtaining empirical evidence for AAT and AAA and stated: “Advocates of programs using animals have begun to see empirical documentation as a critical factor in widespread acceptance of animal-assisted activity (visitation programs) and animalassisted therapy (programs with specific goals for individuals) as beneficial interventions for patients” (pp. 422–423). They cautioned that “...better designed experiments are needed to enhance the credibility of animal-assisted interventions...” (p. 438). Since then there have been many published peer-reviewed studies of AAT and AAA buttressing claims of therapeutic efficacy for these and related activities. However, it is critical to move from collecting basic data on efficacy to refining our experimental methods in order to better understand the nature of these findings. In other words, we currently need to probe the various components of validity in AAT and AAA. More recently, in a review of the effects of pets on human health, Herzog (2011) stated: “Design problems are common in studies of human–animal interactions” and “ ... it is often difficult or impossible to eliminate placebo effects via traditional methods such as singleand double-blind experimental and control groups” (p. 238). Although Herzog’s comments were made specifically about studies of the health effects of pets, they are relevant to the current literature on AAT and AAA. In order to explore the current status of AAT and AAA, I will examine two meta-analyses as well as individual studies published in peer-reviewed journals since 1995 and will focus upon issues of validity and on construct validity in particular. Review of Meta-Analytic Studies In 2007, two major meta-analytic reviews of animal-assisted therapy were published. One, by Nimer and Lundahl (2007), was a comprehensive evaluation of AAT studies from 1973 to 2004, yielding 49 studies that met criteria for inclusion in the meta-analysis. The review covered a wide range of animals, disorders, therapeutic settings, client age categories, and outcome measures. The other, Souter and Miller (2007), consisted of an initial review of 60 studies with only five meeting the inclusion criteria for meta-analysis. Moreover, Souter and Miller focused exclusively on the effect of AAT/AAA on depression. Taken together, both meta-analyses found moderate positive effects for a range of disorders. Nimer and Lundahl did not exclude studies lacking a control group but they found similar results for uncontrolled and controlled studies in their meta-analysis. However, they noted that because AAT is almost always used as an adjuvant to standard therapies it is very difficult to gain a “universal understanding” (Nimer and Lundahl, p. 235) of exactly what AAT is and what components of it might be effective. Nimer and Lundahl (2007) and Souter and Miller (2007) showed that many peer-reviewed and published AAT and AAA studies do not meet minimal standards of research design and therefore cannot be employed in a quantitative meta-analysis. Nimer and Lundahl included studies that lacked a proper control group and still found that only 49 of 250 published studies were eligible for inclusion on the basis of simple criteria such as adequate sample size and enough data to calculate effect sizes. Although finding moderate positive effects for depression, Souter and Miller reported that only one study employed AAT strictly, precluding them from differentiating the effects of AAT from AAA and other general procedures involving animals. Moreover, only one study controlled for the role of the human in the animal condition. That is, only one study of the five included a “human only” control group to be compared with the “human and animal” group. Therefore, the role of the animal in the therapies was not investigated in any systematic way. Therefore, to summarize the two 2007 meta-analyses discussed above, there exist numerous unresolved methodological issues in the AAT and AAA literature. Foremost among these is the lack of research specifically addressing the degree to which the purported positive effects of AAA and AAT are attributable to contact with the human facilitating the animal interaction. Most studies do not include a way to differentiate the animal from the human in AAT and AAA.